IO Serialization¶

Schemas can be serialized to YAML or JSON for storage, version control, or sharing.

YAML¶

import numpy as np
import pandera.xarray as pa

schema = pa.DataArraySchema(
    dtype=np.float64,
    dims=("x", "y"),
    name="temperature",
    nullable=False,
    coerce=True,
)

yaml_str = pa.to_yaml(schema)
print(yaml_str)

schema_type: data_array
dtype: <class 'numpy.float64'>
dims:
- x
- y
name: temperature
coerce: true
coords: null

Restore the schema from YAML:

restored = pa.from_yaml(yaml_str)
print(type(restored).__name__, restored.name)

DataArraySchema temperature

JSON¶

json_str = pa.to_json(schema)
restored = pa.from_json(json_str)
print(type(restored).__name__, restored.name)

DataArraySchema temperature

Both formats support DatasetSchema as well:

ds_schema = pa.DatasetSchema(
    data_vars={
        "temp": pa.DataVar(dtype="float64", dims=("x",)),
    },
    coords={"x": pa.Coordinate(dtype="float64")},
)
json_str = pa.to_json(ds_schema)
restored = pa.from_json(json_str)
print(type(restored).__name__)
print(list(restored.data_vars.keys()))

DatasetSchema
['temp']

File round-trip¶

Write to and read from files:

pa.to_yaml(schema, stream="schema.yaml")
schema = pa.from_yaml("schema.yaml")

pa.to_json(schema, target="schema.json")
schema = pa.from_json("schema.json")